To the Editor.
—We read with interest the article by Frasure-Smith et al1 but noted problems that warrant comment. All of the confidence intervals (CIs) for the hazard ratios (HRs) are incorrect. We noted that eight risk factors in Table 2 had negative lower limits that are outside the range for a ratio. Second, the category daily smoker had lower limits greater than 1.0 with a P value reported to be.17. Since the 95% CI did not include 1.0, the HR is, by definition, statistically significant at the .05 level. Third, the reported CIs are symmetric around the HR, which will not occur since the value of 1.0 defines no effect and the minimum value must be positive.The findings and interpretation of the multivariate analysis are also suspect. The bivariate analysis showed five factors besides major depression with HRs greater than 3.0. Each of these factors had a relatively
Wei M, Hornung CA. Depression and Survival Following Myocardial Infarction. JAMA. 1994;271(14):1080. doi:10.1001/jama.1994.03510380036027